The snail gene required for mesoderm formation in ... - Semantic Scholar

1 downloads 0 Views 2MB Size Report
AUDREY ALBERGA1*, JEAN-LOUIS BOULAY11, ELISABETH KEMPE2, CHRISTINE. DENNEFELD1 and MARC HAENLIN2'$. lLaboratoire de Gtnetique ...
Development 111, 983-992 (1991) Printed in Great Britain © The Company of Biologists Limited 1991

983

The snail gene required for mesoderm formation in Drosophila is expressed dynamically in derivatives of all three germ layers AUDREY ALBERGA 1 *, JEAN-LOUIS BOULAY 1 1, ELISABETH KEMPE2, CHRISTINE DENNEFELD1 and MARC HAENLIN2'$ l Laboratoire de Gtnetique MoUculaire des Eucaryotes du Centre National de la Recherche Scientifique, Unili 184 de Biologic MoUculaire et de Ginie Gtnttique de I'INSERM, Institut de Chimie Biologique, Faculte' de Medicine, 11 rue Humann, 67085 Strasbourg Cedex, France 2 lnstitutfur Entwicklungsphysiologie, Umversit&t zu Kdln, Gyrhofstrasse 17, 5000 Ko'ln 41, Federal Republic of Germany

* Author for correspondence t Present address: Laboratory of Immunology, National Institute of Allergy and Infectious diseases, National Institutes of Health, Bethesda MD 20892, USA t Present address: LGME-CNRS and INSERM-U184, 11 rue Humann, 67085 Strasbourg Cedex, France

Summary The zygotic effect gene snail (sna) encodes a zinc-finger protein required for mesoderm formation in Drosophila embryos. By in situ analysis, sna transcripts are first detected at syncytial blastoderm and persist until very late stages of embryogenesis. Expression of sna is transient and is observed in tissues derived from all three germ layers. Prior to germband elongation, sna RNA accumulation is consistent with its genetically determined role in mesoderm formation. Starting at germband elongation, a second phase of sna expression appears to be initiated, characterized by a highly

dynamic accumulation of transcripts in the developing central and peripheral nervous systems. Translation of sna RNA is apparently delayed as the sna protein is not detected before the onset of gastrulation. Its regional distribution generally correlates with that of sna transcripts. The complex pattern of sna expression strongly suggests that the function of the gene is not restricted to mesoderm formation.

Introduction

mesoderm that will give rise to musculature and other mesodermal derivatives. Ventrolateral cells will give rise to the neuroectoderm that will form the ventral nerve cord and ventral epidermis. Dorsolateral cells will form the peripheral nervous system and dorsal epidermis and extreme dorsal cells will give rise to the amnioserosa. Approximately twenty maternally and zygotically active genes required for the dorsal ventral pattern have been identified (Anderson and NiissleinVolhard, 1984; Anderson, 1987). Positional information along the dorsoventral axis is provided by the action of twelve maternal effect genes, eleven of which make up the 'dorsal group'. Null mutations in any one of the genes in this group result in a dorsalising phenotype whereby all embryonic cells follow a dorsal developmental pathway (Anderson and Niisslein-Volhard, 1984). The current hypothesis is that the product of one of the 'dorsal group' genes, dorsal (dt), acts as the morphogen that determines dorsal-ventral pattern by selectively activating or repressing the expression of downstream zygotic genes (Rushlow et al. 1989; Steward, 1989; Roth et al. 1989). The positiondependent formation of the diverse cell types therefore

The genetic control of Drosophila embryogenesis involves the interaction of both maternal and zygotic information. Two independent systems provide the positional information required for correct pattern formation in the embryo. One includes the genes controlling the anterior-posterior pattern and the other the genes controlling the dorsal-ventral pattern (Niisslein-Volhard, 1979). In each case, maternal effect genes first define the spatial coordinates along the axis. Subsequently, the maternally generated information is interpreted by the zygotic genome, resulting in the spatially restricted expression of genes along the two axes. Whereas the anterior-posterior pattern is a repetition of metameric units, the dorsal-ventral pattern is a nonrepetitive series of cell types (Anderson 1987 for review). For cells to know if they are to differentiate into epidermis, gut, neurons or muscles they must be provided with information as to their position along the dorsal-ventral axis. Cells surrounding the ventral midline will invaginate at gastrulation to form the

Key words: Drosophila, snail gene, finger protein, embryogenesis, mesoderm formation, germ layers.

984

A. Alberga and others

depends upon the activity of several zygotically active genes, which are selectively expressed in response to the maternal factors. Null mutations in two of these zygotic genes, sna and twi, result in the lack of normal ventral furrow formation with the subsequent absence of mesodermal structures (Grau et al. 1984). Although the phenotypes of the two mutants are similar, they are not identical. Strong sna mutant alleles show abnormal development of laterally derived ectodermal structures as well as the ventrally derived mesoderm, whereas in the weaker alleles only the most ventrally derived structures are affected. This variation of phenotypes is not seen in twi mutant alleles where only mesodermal derived structures are affected (Simpson, 1983). The twi gene is expressed in the midventral region of the embryo, corresponding to the presumptive mesoderm and encodes a protein that contains a putative DNAbinding helix-loop-helix motif (Thisse et al. 1988). The previous characterization of sna (Boulay et al. 1987) predicted a 43xlO 3 M r protein with five zincfinger motifs, which suggested that sna may be involved in transcriptional regulation. Positive identification of the gene was obtained by inducing sna phenocopies in wild-type embryos following the injection of antisense RNA. Subsequent P-element rescue experiments showed 13 kb of flanking sequences to be sufficient to rescue the lethality associated with the sna mutant phenotype, with survival to adulthood (Boulay, 1988). We present the results of studies on sna RNA and protein distribution during embryogenesis which show a dynamic expression of the gene in all three germ layers of the developing embryo. The complex pattern of expression suggests that sna may be required in a large range of tissues and implies that the function of the gene is not restricted to mesoderm formation. Finally, we present data showing that prior to invagination of the mesoderm, sna is expressed in twi~ embryos, and twi is expressed in sna~~ embryos, suggesting distinct roles for the two genes in the process of mesoderm formation. Materials and methods Fly strains Canton S and Oregon R raised at 25 °C on standard com meal medium were used as our wild-type reference stocks. The snail allele, sna1**1, and the twist allele, twiEY5°, were provided by P. Simpson. Fusion protein and antisera Two fragments, one of 717 bp, which codes for amino acids 151-390 (FP-pBl), and the other of 441 bp, encoding amino acids 243-390 (FP-pCl), were excised from plasmids containing DNAasel-generated deletions of the cDNA clone pcSB (see Boulay et al. 1987). A third fragment of 1185 bp, which encodes the complete protein (FP-pD12), was obtained after site directed mutagenesis to introduce a Hin&Wl site 15 nt upstream of the first ATG. HindlU fragments of the cDNA were inserted in frame (as well as in the antisense orientation) in the procaryotic vector pUR291 (Riither and Muller-Hill, 1983) to create lacZ-sna fusion genes, which were then introduced into E. coli by transformation. Expression was induced by IPTG and fi-gal-sna fusion proteins were

prepared essentially according to Rio et al. (1986). For preparations used in immunization, bacteria pellets from induced cultures were lysed in SDS sample buffer (Laemmli, 1970) and loaded onto a preparative polyacrylamide-SDS gel. After electrophoresis, the gel was stained with Coomassie blue, the fusion protein band was excised and pulverized in liquid nitrogen. Alternatively, protein aggregate pellets were suspended in 8M urea, 50mM Tris-HCl pH7.5, 500mM NaQ, lmM EDTA, 50 mM DTT, 10 fm ZnSO4, lmM PMSF, antiprotease cocktail (10 /m TPCK, 10 /JM TLCK, 0.15 JZM pepstatin, 0.1 /MA leupeptin), extracted by mild magnetic stirring for 45min at 0-4 °C and centrifuged for lOmin at 10 000 revs min"1. The extracted proteins were renatured by the stepwise removal of urea after dialysis against solutions of 5 M urea, 50 mM Tris-HCl pH7.5, 200mM NaCl, 20mM DTT, 10 /an ZnSO4, 0.2 mM PMSF, anti-protease cocktail, 10% glycerol followed by 2 M urea in the same buffer and finally against 50mM Tris-HCl pH7.5, 200 mM NaCl, 2mM DTT, 10/XM ZnSO4, 10% glycerol. Insoluble material was removed by centrifugation at 10000revsmin"1 for lOmin. These preparations were primarily used for preparing fi-gal-sna affinity resins (pBl and pD12) by coupling the proteins to CNBr-activated Sepharose, according to the manufacturer's (Pharmacia) protocol. Antisera An equal volume of complete Freund's adjuvant was added to gel-purified fusion protein (pBl or pD12) suspended in PBS and approximately 200/xg of fusion protein was injected subcutaneously (at several sites) into female rabbits. The animals were boosted after 2 weeks with the same amount of protein in incomplete adjuvant and again after 4 weeks with 100 ng of fusion protein in PBS and were bled 10 days following the second booster. Thereafter the animals were periodically boosted with fusion protein in PBS, bled 10 days later and the titer of the serum was monitored by western blot analysis. Serum was adsorbed on an affinity column prepared from proteins (from IPTG-induced pUR 291 transfected E. coli) coupled to CNBr-activated Sepharose. Flow-through material from this column was loaded onto a pBl (anti-pBl) or pD12 (anti-pD12) fusion protein affinity column and the antibodies eluted with 50mM glycine-150mM NaCl, pH2.5, neutralized immediately and dialyzed against PBS. The specificity of the serum was confirmed by western blot analysis. Antibody-protein complexes were revealed with 125I Protein A in PBS-3% low-fat milk. All western analyses included appropriate controls, proteins prepared from the vector alone (induced or non-induced), from the non-induced fusion gene plasmid and from plasmids with inserts fused in antisense orientations. The two znti-sna sera gave similar qualitative results. However, as the anti-pBl serum had the higher titer, this preparation was used in immunodetection assays (see below). Protein extracts The sna cDNA, pcSB, cloned in both orientations into the ZscoRI site of the eukaryotic expression vector pSG5 (Green et al. 1988) was introduced into HeLa cells by calcium phosphate transfection (Ausubel et al. 1987). The transfected cells were lysed by three freeze-thaw cycles in TEBG 500 (see Bocquel et al. 1989), the proteins separated on 10% SDS-PAGE and transferred to BA85 nitrocellulose membranes (Schleicher and Schiill). In parallel, proteins were extracted from dechorionated unstaged wild-type embryos essentially as in Soeller et al. (1988), separated on 10% SDS-PAGE and transferred to nitrocellulose membranes.

Dynamic expression of sna during embryogenesis Immunostaining Embryos were dechorionated with 50% household bleach, fixed with 4% paraformaldehyde-PBS/heptane (v/v), devitellinized in heptane/methanol-25 HIM EGTA (V/V), rehydrated in PBS-0.1 % Tween 80 (PBT) and blocked with 5 % decomplemented normal goat serum (NGS)-PBT. Embryos were incubated with affinity-purified serum overnight at 4 C and washed with 5 % NGS-PBT. Incubation with the second antibody, aLkaline-phosphatase-conjugated anti-rabbit IgG (1/1000 dilution), was at room temperature during 2h and was followed by extensive washings with PBT and with 100HTM Tris-HCl pH9.5, 100mM NaCl, 50am MgCl2 (coloring buffer). The color was developed by adding 4.5 jA of nitroblue tetrazolium salt (TSmgml"1, 70% dimethylformamide), 3.5 (A of 5-bromo-4-chloro-3-indolyl phosphate (SOmgmT1, dimethylformamide) to the last ml of coloring buffer wash and incubating for the appropriate period (5-15 min). The reaction was stopped by the addition of 10mM Tris-HCl pH8-lmM EDTA (TE), embryos were washed with PBS and mounted in 60% glycerol-PBS.

In situ hybridization Antisense sna mRNA (for preparation, see Boulay et al. 1987) was labeled with 35S-UTP and hybridized to paraffinsectioned embryos as previously described (Ingham et al. 1985). Digoxigenin labeled DNA was synthesized according to the Boehringer-Mannheim protocol and hybridized to whole-mount embryos according to Tautz and Pfeifie (1989). Results

(A) Localisation of sna transcripts during embryogenesis Transient accumulation of sna RNA in the mesoderm In situ hybridization to sectioned and whole-mount embryos was used to determine the spatial distribution of sna RNA during embryogenesis. The probes used were 35S-labeled antisense sna RNA and a digoxigeninlabeled 547 bp Hindlll fragment of the cDNA, which includes the entire finger coding region. The sna transcripts are first detected during late syncytial blastoderm (stage 4) when trace amounts of RNA present at a ventroposterior position are revealed with the 35S probe (not shown). Subsequently, there is an increase in the level of the transcripts and sna RNA is readily detected along the ventral surface of the embryo extending into the anterior and posterior poles (Fig. 1A). In the period between late syncytial and early cellular blastoderm, the level of sna transcripts in the posterior region begins to decrease (Fig. IB) and at the completion of cellular blastoderm (stage 5), sna RNA is no longer detected in the posterior region of the embryo. Thus, in the period immediately preceding gastrulation (late stage 5 to early stage 6), sna RNA is present in the anlagen of the anterior midgut and mesoderm (Fig. 1C). At the onset of gastrulation, uninvaginated cells of the presumptive mesoderm are intensely labeled (Figs 1D,2B). As gastrulation proceeds, the strong signal is maintained as the cells invaginate (Fig. 2B), whereas there is a progressive decrease in the levels of

985

sna RNA in the invaginated cells (Fig. 2C; compare Fig. ID and IE). By the end of gastrulation, only low levels of sna RNA are detected in the invaginated mesoderm while high levels are now seen in cells of the mesectoderm (Fig. IF), an accumulation that lasts until stage 7. The low levels of sna transcripts present in the mesoderm persist until stage 8 (Fig. 1G). Dynamic accumulation of sna transcripts in the developing nervous system The most dynamic phase of sna expression occurs during germ band elongation, as shown by a rapidly changing pattern of RNA accumulation. In the early phase of germ band elongation (late stage 6), sna RNA is seen in the region of the anterior midgut rudiment as well as in a region on the dorsal surface of the invaginating amnioproctodeum. As germ band extension progresses (stages 7-9), sna RNA persists in the anterior midgut until stage 9 at which time transcripts are barely detectable in this tissue (Fig. 1H). Starting at late stage 8, sna RNA begins to accumulate in the neuroectoderm (Fig. 1G) particularly in large cells that probably correspond to neuroblast precursors and to segregated neuroblasts (Fig. 1H). In addition, sna RNA is present in cells lying between the proctodeum and the posterior midgut which may correspond to the primordium of the Malpighian tubules (Fig. 1H). During the slow phase of germ band extension (stages 9 to 10), sna transcripts are evident in segregating neuroblasts of the germ band and in procephalic neuroblasts (Fig. 1H,I). The accumulation in the neuroblasts is not uniform and, within the developing nervous system, sna transcripts are detected in some but not in all neuroblasts of an embryo (Fig. 1J,K). However, because of the highly dynamic expression one cannot exclude that sna is expressed in all neuroblasts, but not simultaneously. The sna transcripts persist in the neuroblasts until stage 11 when they are barely detected. At this time, they are also present in the median cells (Fig. 1M) and persist until stage 14. After germband retraction, staining is again seen in the central nervous system and lasts until stage 16. sna RNA persists until late embryogenesis Transient accumulation of sna transcripts is observed in several groups of cells, some of which we have tentatively identified from their time of appearance, position and morphology. Around stage 10, sna RNA appears in laterally positioned individual cells (Fig. IK) that from the time of their appearance, presumably correspond to precursor cells of the peripheral nervous system (Ghysen and O'Kane, 1989). At the start of germband retraction (late stage 11), the signal in these cells is weaker and sna RNA begins to accumulate in a group of cells in each of the first seven abdominal segments (Fig. 1L). On the basis of their number and of their location, these cells probably correspond to progenitor cells of the pentascolopidial chordotonal sensory organs (Ich5, see Campos-Ortega and Hartenstein, 1985). The accumulation in these organs will continue until stage 14. During stage 13, sna RNA

986

A. Alberga and others

nb

nb

K

pPNS

M

me

N

Ich5

O

ape

Ich5

wd

hd

Fig. 1. Pattern of sna expression in wild-type embryos. Whole-mount embryos at various stages of development were hybridized with a digoxigenin-labeled 547 bp Hindlll fragment from the cDNA. For all embryos, anterior is to the left and dorsal is up (except E, F, J, M, O). Staging is according to Campos-Ortega and Hartenstein (1985). Embryos were photographed using Nomarski optics. Arrows indicate labeled structures or landmarks. (A) Syncytial blastoderm embryo (stage 4). Arrowheads indicate the anterior and posterior limits of the region of expression, note reduction in early stage 5 embryo (B) and in late stage 5 embryo (C); pole cells (pc). (D) Early gastrulating embryo (stage 6), invaginating mesoderm (ims), border ventral furrow (bvf). (E) Ventral view of late gastrulating embryo (early stage 7), anterior midgut (am), cephalic furrow (cf); cells on the border of the ventral furrow (vf) are about to invaginate. (F) Detail of the ventral furrow region at the end of gastrulation (stage 7, embryo slightly older than in E). Note stronger signal in the mesectoderm (mec) than in the mesoderm (ms). (G) Stage 8 embryo (germband extension) transcripts start to accumulate in the ectoderm (ec). (H) Stage 9 embryo, neuroblasts (nb), Malpighian tubules (mt). (I-K) Stage 10 embryos (extended germband) strong signal in the neurogenic region (J, ventral view); stomodeum (sto), presumed precursor cells of the peripheral nervous system (pPNS). (L,M) Lateral and ventral views, respectively, of a stage 11 embryo, Pentascolopidial chordotonal (Ich5) and Bolwig organ precursors (Bo) and median cells (me) are indicated. (N) Germband retracted embryo (stage 13), presumed precursors of wing disc (wd) and haltere disc (hd). (O) Embryo at the end of embryogenesis (stage 16—17), cells in the region of the anal plate (ape).

Dynamic expression of sna during embryogenesis

987

B

Fig. 2. Distribution of sna transcripts during mesoderm formation. A 35S labeled antisense RNA probe was hybridized to 8/an cross sections through gastrulating embryos. (A) End of stage 5, strong signal in the anlage of the mesoderm. (B) Stage 6, start of mesoderm invagination. (C) later stage 6, note the reduced signal in invaginated mesoderm.

accumulates in two clusters of cells of the mesothoracic and metathoracic segments, which probably correspond to precursors of the wing and haltere imaginal discs (Fig. IN). Transcripts are also detected in cells that from their position and morphology, presumably correspond to Bolwig photoreceptor organs (Bolwig, 1946). At the end of embryogenesis (stages 16-17), some sna RNA is still present in the presumed wing and haltere disc precursors as well as in a group of cells located in the vicinity of the anal plate (Fig. 1O). (B) Distribution of sna protein We analyzed the distribution of sna protein during embryonic development by staining whole-mount embryos with antibodies raised against lacZ-sna hybrid proteins. All immunodetection analyses were performed with monospecific antibodies prepared by affinity chromatography (see Materials and Methods). These antibodies recognized a single protein band with an apparent relative molecular mass of approximately 48X103 in wild-type embryonic extracts as well as in extracts of HeLa cells transiently expressing the sna protein from the eukaryotic expression vector pSG5

M

sna •

Emb Emb

»••» ^Hfr

M

^

(Green et al. 1988) but did not recognize any protein produced when the coding region was inserted in pSG5 in the antisense orientation (Fig. 3). The signal obtained with the antiserum was either abolished or significantly lowered if the serum was preincubated with fusion protein (2h at 25 °C) before incubation with embryos or with nitrocellulose transfers of protein extracts. Finally, no specific staining was detected with the antiserum in homozygous sna embryos (not shown). Whereas sna transcripts first appear during syncytial blastoderm (stage 4), the sna protein is first detected at the onset of gastrulation (stage 6); thus there is a significant lag in the appearance of the translation product. At the start of gastrulation, weak staining of sna protein is observed in individual cells along the ventral surface (Fig. 4A) and shortly afterwards in the primordium of the anterior midgut and in the mesoderm (Fig. 4B). The protein is detected only in a subset of the invaginated cells and is apparently located in the nucleus (Fig. 4B,C). During the early phases of ventral furrow formation, cells within the furrow are more intensely stained than those bordering the furrow (Fig. 4C,D), which is in contrast to the RNA accumu-

p131 p131 p13il

^^^»

Fig. 3. Monospecific sna antiserum recognizes recombinant and endogenous sna protein. Western blots of extracts of wild-type embryos (Emb) and of HeLa cells transiently transfected with pSG5 based sna expression vectors with sna cDNA inserted in the sense orientation (pl31) and in the antisense orientation, (pl31I). Two separate preparations of Emb and pl31 were analyzed. See Material and methods for preparation of protein extracts and immunodetection procedure.

988

A. Alberga and others

Fig. 4. Localisation of sna protein in wild-type embryos. All embryos were incubated with affinity-purified anti-/S-ga/-57W protein antiserum (pBl). Anterior is to the left and dorsal is up (except D,E,H,J). Staging is according to Campos-Ortega and Hartenstein (1985) and photographs were taken with Nomarski optics. (A) Detail of the ventral furrow at the onset of gastmlation (stage 6). Arrowhead indicates a cell expressing the protein. (B) lateral view of an early gastrulating embryo (stage 6). Arrowheads indicate the limits of the region of expression, cephalic furrow (cf). (C) Detail of the ventral furrow of a mid gastrulating embryo (stage 6). Arrowhead indicates a non-expressing cell. (D) Ventral view of a mid gastrulating embryo (stage 6), arrowhead indicates less intensely stained cells on the border of the furrow. (E) Ventral view of a late gastrulating embryo (slightly older than in D), cells on the border of the furrow are more intensely stained (arrowhead) than cells inside the furrow. (F) Lateral view of embryo during early germband elongation (stage 7, focus is on the dorsal folds), cephalic furrow (cf), anterior transversal fold (atf), posterior transversal fold (ptf), lateral proctodeal fold (lpf). (G) Stage 8 embryo (germband elongation). Low protein levels are detected in the mesoderm (ms), anterior midgut (am), ectoderm (ec). (Note that staining was extended deliberately.) (H) Dorsal view of a stage 9 embryo, neuroectodermal cells are stained. (I) Lateral view of embryo at about stage 10-11, showing staining in presumed precursor cells of the peripheral nervous system (pPNS). (J) Ventral view of a stage 11 embryo, neuroectodermal cells (as in H) are stained for the protein.

lation (see above). However, at the end of gastrulation, cells located on the border of the ventral furrow become more strongly stained for protein than cells of the invaginated mesoderm. Thus, at this stage of embryogenesis, the distribution of the protein coincides with the accumulation pattern of sna transcripts (compare Figs 2E and 4E). The regional distribution of the protein generally correlates with that of transcripts; however, with the start of germband elongation, some differences are observed. During the early phase of germband elongation (stage 7), weak staining of the protein is seen in the regions of the cephalic furrow and transversal folds (Fig. 4F). In later phases of germband elongation, the intensity of the signal increases and, in

the neurogenic region, the protein is detected only in a subset of the ectodermal cells (Fig. 4H-J). However, as the distribution pattern of the protein does not correspond to that seen during the wave of neuroblast segregation (Hartenstein and Campos-Ortega, 1984), it is not known if all cells that express sna protein will become neuroblasts or if all neuroblasts express the protein. Surprisingly, in the neurogenic region, staining is apparently restricted to the cytoplasm in the majority of cells (Fig. 4H-J). During the later stages of embryogenesis (until stages 16-17), only low levels of the protein are detected (data not shown). This late expression includes the presumptive wing and haltere disc precursors.

Dynamic expression of sna during embryogenesis (C) sna expression in twi embryos and twi expression in sna" embryos The two zygotic genes, sna and twi, occupy similar positions in the regulatory hierarchy required for the development of the mesoderm (reviewed by Anderson, 1987) and may act sequentially with one gene controlling the other. To investigate this possibility, we looked at the expression of twi protein in sna embryos by immunodetection using antibodies directed against the twi protein (Thisse et al. 1988). At blastoderm, there is no detectable difference in the distribution of twi protein between heterozygous ('wild-type') and homozygous sna embryos. The first visible difference between the two types of embryos is seen at the end of gastrulation (stage 6-7). In the heterozygous embryo, fwi-positive cells invaginate (Fig. 5A,C) whereas in a homozygous sna embryo, these cells do not invaginate and remain at the surface (Fig. 5B,D). During germband elongation (late stage 8), twi expression decreases and only a few ftvi-positive cells remain in the small ventral fold which is formed in sna" embryos (compare

Wt

989

Fig. 5E,F). Thus it would appear that, prior to gastrulation, sna is not essential for twi expression. As the sna protein is not detected before the start of gastrulation, the expression of sna in early twi mutant embryos was analyzed by the distribution of sna transcripts. The results show that, at blastoderm, sna transcripts are present in twi~ embryos. The expression of sna in twi~ embryos and of twi in sna~ embryos, prior to gastrulation, excludes a simple hierarchical relationship between the two zygotic genes for formation of the mesoderm. The spatial expression of sna is more restricted in twi~ embryos and the transcripts accumulate in a narrower region than in wild-type embryos (compare Fig. 6A and B). However, this may be a general effect of a reduction in the size of the mesoderm anlage in twi embryos (see below). Discussion

sna expression during mesoderm formation The first evidence of sna expression is obtained during

sna~

Fig. 5. Expression of twi in wild-type and snaRYI embryos. All embryos were incubated with anti-twi protein antiserum (Thisse et al. 1988). (A,C) early stage 8 wild-type embryo (same embryo, different focal plane); (E), late stage 8 wild-type embryo; (B,D), early stage 8 sna embryo (different planes of focus of the same embryo); (F) late stage 8 snaRY1 embryo. Arrows indicate invaginated hv/-positive cells in A and E and non-invaginated ow-positive cells in B and F. Arrowheads in C and D give limits of fvW-positive cells which remain at the surface.

990

A. Alberga and others

Wt

B

twiFig. 6. Expression of sna transcripts in wild-type and twiEY50 embryos. Whole-mount embryos were hybridized with digoxigenin-labeled sna probe. Ventral views of stage 5 wild-type embryo (A) and stage 5 twiEYS0 embryo (B). Arrowheads indicate limits that define the anlage of the mesoderm.

syncytial blastoderm when low levels of transcripts are detected in the region of the proctodeum anlage. These then accumulate progressively along the ventral surface of the embryo and extend into the posterior and anterior poles. At this developmental stage, the spatial distribution of sna transcripts is similar to that of the dorsal (Steward et al. 1988; Roth et al. 1989) and twist proteins (Thisse et al. 1988). This is not surprising as the expression of all three genes is required for mesoderm formation. It has been shown that the dorsal gene product is the morphogen that controls dorsal-ventral, pattern by activation or repression of zygotic regulatory genes (Rushlow et al. 1989; Steward, 1989; Roth et al. 1989). In addition, Simpson (1983) reported synergistic interactions between dl and twi and dl and sna, suggesting possible interactions between the products of these three genes. At blastoderm, in embryos issued from homozygous dl females, sna transcripts are not detected in the mid-ventral region of the embryo (region of the presumptive mesoderm in wild-type embryos) though low levels of the RNA are detected in other regions of the embryo (A. Alberga, unpublished results). The fact that sna is expressed in twi~ embryos and that twi is expressed in sna~ embryos excludes a simple relationship in which the product of one gene is essential for the expression of the other. Nevertheless, the spatial expression of sna is reduced in the absence of the twi gene product (see Fig. 6B). A comparable

reduction in the size of the mesoderm anlage was also observed at blastoderm, in twi~ embryos hybridized with probes corresponding to the transcripts of E(spl)m8 and Delta (Kempe, Haenlin and CamposOrtega, unpublished results). Therefore, the modification of sna expression may reflect a general effect of a reduction in the size of the mesoderm anlage in twi~ embryos. In contrast, the absence of the sna gene product has no effect on twi expression, prior to gastrulation (see Fig. 5). Together, these results suggest that, if a hierarchial relationship exists between these two genes, then twi is epistatic to sna in the process of mesoderm formation. In the interval between syncytial blastoderm and gastrulation, the spatial expression of sna is consistent with a role in mesoderm formation. During cellular blastoderm, sna expression is apparently turned off in cells of the posterior pole and its transcripts extend anteriorly to include the anlagen of the anterior midgut and mesoderm. Thus immediately preceding the onset of gastrulation, the spatial expression of sna singularizes since during the same period, both dl and twi protein expression continue in the posterior pole (Steward, 1989; Roth et al. 1989 and Thisse et al. 1988, respectively). In the gastrulating embryo, high levels of sna transcripts are present in the presumptive mesoderm before and during invagination whereas only low levels are found in the invaginated tissue. Thus the expression of sna in mesodermal cells is apparently limited to a relatively brief period during the early stages of embryogenesis. In comparison, twi expression is maintained in the mesoderm until much later in embryogenesis (Thisse et al. 1988). The difference in their temporal patterns of expression contributes to the increasing evidence for separate roles of these two genes in the development of the mesoderm (see Leptin and Grunewald, 1990). Although sna transcripts are present at syncytial blastoderm (stage 4), sna protein is not detected before the onset of gastrulation (stage 6). Similar delays between the appearance of RNA and gene products have been reported for other Drosophila developmental genes, e.g. zerkniillt (Rushlow et al. 1987), slit (Rothberg et al. 1988) and single-minded (Crews et al. 1988). While the ventral furrow is being formed, the sna protein is detected in nuclei of cells within the furrow and these are more intensely stained than cells surrounding the furrow. This suggests that during gastrulation the protein is essentially nuclear (see below) and accumulates primarily in cells that are in the process of invagination (see Fig. 4A-D). In comparison, during this period, cells that are about to invaginate have higher levels of sna RNA than cells that are in the process of invagination (compare Figs ID and 4D). In the presumptive mesoderm, sna RNA and twi protein are expressed in the same cells; however, not all fvw-positive cells invaginate during gastrulation (Leptin and Grunewald, 1990). The localization of the sna gene product in invaginating cells suggests that sna may be required for the invagination process of the future mesodermal cells.

Dynamic expression of sna during embryogenes is sna expression is not restricted to the mesoderm If the early expression of sna is consistent with the phenotypes of the mutants, its later expression is somewhat unexpected. Throughout embryogenesis sna transcripts accumulate transiently in a variety of tissues in which there is no known function for the gene. These include the endodermal derived midgut and ectodermal derived structures, in particular, that of the developing nervous system. Around stage 8, sna RNA is first detected in cells from which neuroblasts will segregate and afterwards in the segregating neuroblasts (stage 9). In addition, the transcripts are present during the formation of the peripheral nervous system. The accumulation of sna RNA in the nervous system is very dynamic, which makes it difficult to describe fully its spatial pattern of expression. At present, we have no explanation for the accumulation of sna RNA in the precursors of the nervous system. Preliminary observations on sna mutants (El Messal, 1987) showed that, in sna mutants, the segregation of neuroblasts occurs; however, it is not known if their number and/or their identity are correct. The neural cord formed is often abnormal with ganglion-like structures, its left and right sides often fail to fuse and later to condense into a normal CNS. These defects could be due either to the lack of the sna gene' product or to a secondary mechanical defect, resulting from the absence of mesodermal structures and the twisted nature of the mutant embryo. It remains to be seen if sna is involved in neurogenesis and in the formation of the endoderm. A temperature-sensitive allele would be useful for this analysis. In very late stages of embryogenesis (16-17), sna transcripts are found in the presumed precursors of the wing and haltere discs. This result is of particular interest and it is likely that failure of sna activities in these discs leads to the missing halteres and hemithorax phenotypes seen in both genetic (Grau et al. 1984) and transgenic studies (Boulay, 1988). During germband elongation in the neurogenic region, the protein is accumulated only in a subset of the cells that accumulate RNA. A differential accumulation of transcript and protein was reported for the Drosophila segmentation gene Krilppel (Gaul et al. 1987) and for lethal of scute, one of the genes in the achaete-scute complex (Cabrera, 1990). In both cases, post-transcriptional regulation was invoked as the basis for this difference. As there is a significant delay between the appearance of sna RNA and sna protein, it is possible that post-transcriptional regulation is also involved in the case of the sna protein. Subcellular localisation of the sna protein Although in some cells the protein appears to be in the nucleus, the majority of the cells in the neurogenic region clearly show cytoplasmic localisation of the protein (see Fig. 4H-J). Note that in HeLa cells the transiently expressed protein is nuclear (A. Alberga and T. Ylikomi, unpublished results), which confirms that the protein contains a functional signal for nuclear accumulation. What could be the significance of the cytoplasmic localisation? If the sna protein acts only as

991

a nuclear transcription factor then it will be active during mesoderm formation when the protein is apparently located in the nucleus (see Fig. 4A-C) and to a much lesser extent in the developing nervous system. This would be similar to the case of dorsal where it has been shown that the dl protein is located in the nucleus in regions where the gene is known to have a function and is cytoplasmic where there are no known genetic requirements (Rushlow et al. 1989: Steward, 1989; Roth et al. 1989). It is however, possible that the cytoplasmic sna protein has a distinct function. By virtue of its zincfingers,it shows structural homology to Xenopus TFIIIA transcription factor and we note that TFIIIA binds to the internal control region of 5S RNA genes (Engelke et al. 1980) as well as to the gene product, 5S RNA (Picard and Wegnez, 1979). Klug and Rhodes (1987) suggested that the ability to bind to both DNA and RNA could be a general property of TFIIIAlike finger proteins (i.e. bearing the Cys-Cys . . . . FlisFlis motif), thus providing the possibility for an additional mechanism for regulation. It is therefore not excluded that the sna protein could be binding to RNA and it remains to be seen if the cytoplasmic localisation is indeed indicative of a novel regulatory role of the sna protein. In spite of this difference in the localisation of the protein in early and late embryogenesis, is there a unifying feature of sna expression? It is noteworthy that in both the early and late phase, sna is expressed in regions that undergo a morphogenetic movement and that the expression apparently ceases once the movement is completed. This suggests a possible role of sna in the process of cell movement. We are grateful to C. Schweitzer for her contribution during the early phase of the preparation of fusion proteins, to T. Ylikomi for kindly performing the subcellular localisation assays, to F. Perrin-Schmitt for the twist protein antibodies and to M. El Messal for allowing us to use her unpublished data. We thank M. Bourouis, J.A. Campos-Ortega, H. Gronemeyer and G. Richards for their critical reading of the manuscript. M.H. who was an EMBO fellow on leave from CNRS thanks Professor Campos-Ortega for his encouragement and helpful discussions. A.A. is grateful to H. Gronemeyer and G. Richards for their advice and fruitful discussions. We greatly appreciate the efforts of those involved in preparing the manuscript: B. Boulay, F. Haenel, J.M. Lafontaine, A. Landmann and C. Werle. This work was supported by grants from the CNRS, the INSERM and from the Deutsche Forschungsgemeinschaft (DFG, SFB 243). References ANDERSON, K. V. (1987). Dorsal-ventral embryonic pattern genes of Drosophila. Trends Genet. 3, 91-97. ANDERSON, K. V. AND NOSSLEIN-VOLHARD, C. (1984). Genetic

analysis of dorsal-ventral embryonic pattern in Drosophila. In Pattern Formation: A Primer in Developmental Biology (ed. G. M. Malacinski and S. V. Bryant), pp. 269-289. New York: MacMillan. AUSUBEL, F. M., BRENT, R., KINGSTON, R. E., MOORE, D. D . , SEIDMAN, J. G., SMITH, J. A. AND STRUHL, K. (1987). Current

Protocols in Molecular Biology. New York: John Wiley and Sons.

992

A. Alberga and others

BOCQUEL, M. T., STRICKER, C.'CHAMBON, P. AND GRONEMEYER,

H. (1989). The contribution of the N- and C-terminal regions of steroid receptors to activation of transcription is both receptor and cell specific. Nucl. Acids Res. 17, 2581-2595. BOLWIG, N. (1946). Senses and sense organs of the anterior end of the houseflylarva. Vidensk. Medd. dansk. naturh. Forenh. Kbh. 109, 80-212. BOULAY, J. L. (1988). Clonage et etude moleculaire du gene snail, intervenant dans la mise en place de l'axe dorso-ventral de 1'embryon de Drosophila melanogaster. PhD thesis. University Louis Pasteur, Strasbourg. BOULAY, J. L., DENNEFELD, C. AND ALBERGA, A. (1987). The

Drosophila developmental gene snail encodes a protein with nucleic acid binding fingers. Nature 330, 395-398. CABRERA, C. V. (1990). Lateral inhibition and cell fate during neurogenesis in Drosophila: the interactions between scute, Notch and Delta. Development 109, 733-742. CAMPOS-ORTEGA, J. A. AND HARTENSTEIN, V. (1985). The

Embryonic Development of Drosophila melanogaster. Berlin: Springer Verlag. CREWS, S. T., THOMAS, J. B. AND GOODMAN, C. S. (1988). The

Drosophila single-minded gene encodes a nuclear protein with sequence similarity to the per gene product. Cell 52, 143-151. EL MESSAL, M. (1987). Etude morphog^netique de trois genes impliqu6s dans la definition du pattern dorso-ventral de 1'embryon de Drosophila melanogaster: twist, snail and shaggy. PhD thesis. Universite Louis Pasteur, Strasbourg. ENGELKE, D. R., NG, S.-Y., SHASTRY, B. S. AND ROEDER, R. G.

(1980). Specific interaction of a purified transcription factor with an internal control region of 5S RNA genes. Cell 19, 717-728. GAUL, U., SEIFERT, E., SCHUH, R. AND JACKLE, H. (1987).

Analysis of Krilppel protein distribution during early Drosophila development reveals posttranscriptional regulation. Cell 50, 639-647. GHYSEN, A. AND O'KANE, C. (1989). Neural enhancer-like elements as specific cell markers in Drosophila. Development 105, 35-52. GRAU, Y., CARTERET, C. AND SIMPSON, P. (1984). Mutations and

chromosomal rearrangements affecting the expression of snail, a gene involved in embryonic patterning in Drosophila melanogaster. Genetics 108, 347-360. GREEN, S., ISSEMANN, I. AND SHEER, E. (1988). A versatile in vivo

and in vitro eukaryotic expression vector for protein engineering. Nucl. Acids Res. 16, 369. HARTENSTEIN, V. AND CAMPOS-ORTEGA, J. A. (1984). Early

neurogenesis in wild-type Drosophila melanogaster. Wilhelm Roux's Arch, devl Biol. 193, 308-325. INGHAM, P. W., HOWARD, K. R. AND ISH-HOROWICZ, D. (1985).

Transcription pattern of the Drosophila segmentation gene hairy. Nature 318, 439-445. KLUG, A. AND RHODES, D. (1987). 'Zinc fingers': a novel protein motif for nucleic acid recognition. TIBS 12, 464—469. LAEMMU, U. K. (1970). Cleavage of the structural proteins during assembly of the head of bacteriophage T4. Nature 227, 680-685.

LEPTIN, M. AND GRUNEWALD, B. (1990). Cell shape changes during gastrulation in Drosophila. Development 110, 73-84. NOSSLEIN-VOLHARD, C. (1979). Maternal effect mutations that alter the spatial coordinates of the embryo of Drosophila melanogaster. In Determinants of Spatial Organization (ed. S. Subtelny and I. R. Koenigsberg), pp. 185-211. New York: Academic Press. PICARD, B. AND WEGNEZ, M. (1979). Isolation of a 7S-particle from Xenopus laevis oocytes: a 5S RNA protein complex. Proc. natn. Acad. Sci. U.S.A 76, 241-245. Rio, D. C , LASKI, F. A. AND RUBIN, G. M. (1986). Identification

and immunochemical analysis of biologically active Drosophila. element transposase. Cell 44, 21-32. ROTH, S., STEIN, D. AND NOSSLEIN-VOLHARD, C. (1989). A

gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo. Cell 59, 1189-1202. ROTHBERG, J. M., HARTLEY, D. A., WALTHER, Z. AND ARTAVANIS-

TSAKONAS, S. (1988). slit: An EGF-homologous locus of D. melanogaster involved in the development of the embryonic central nervous system. Cell 55, 1047-1059. RuTHEit, U. AND MOLLER-HILL, B. (1983). Easy identification of cDNA clones. EMBO J. 2, 1791-1794. RUSHLOW, C. A., HAN, K., MANLEY, J. AND LEVINE, M. (1989).

The graded distribution of the dorsal morphogen is initiated by selective nuclear transport in Drosophila. Cell 59, 1165-1177. RUSHLOW, C , FRASCH, M., DOYLE, H. AND LEVINE, M. (1987).

Maternal regulation of zerkntillt: a homoeobox gene controlling differentiation of dorsal tissues in Drosophila. Nature 330, 583-586. SIMPSON, P. (1983). Maternal-zygotic gene interactions during formation of the dorsoventral pattern in Drosophila embryos. Genetics 105, 615-632. SOELLER, W. C , POOLE, S. J. AND KORNBERG, T. (1988). In vitro

transcription of the Drosophila engrailed gene. Genes Dev. 2, 68-81. STEWARD, R. (1989). Relocalization of the dorsal protein from the cytoplasm to the nucleus correlates with its function. Cell 59, 1179-1188. STEWARD, R., ZUSMAN, S. B., HUANG, L. H. AND SCHEDL, P.

(1988). The dorsal protein is distributed in a gradient in early Drosophila embryos. Cell 55, 487-495. TAUTZ, D. AND PFEIFLE, C. (1989). A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98, 81-85. THTSSE, B., STOETZEL, C , GOROSTIZA-THISSE, C. AND PERRIN-

SCHMITT, F. (1988). Sequence of the twist gene and nuclear localization of its protein in endomesodermal cells of early Drosophila embryos. EMBO J. 7, 2175-2183.

{Accepted 20 December 1990)